Abstract
Background COVID-19 has strained population breast mammography screening programs that aim to diagnose and treat breast cancers earlier. As the pandemic has affected countries differently, we aimed to quantify changes in breast screening volume and uptake during the first year of the COVID-19 pandemic.
Methods We systematically searched Medline, the WHO (World Health Organization) COVID-19 database, and governmental databases. Studies covering January 2020 to March 2022 were included. We extracted and analyzed data regarding study methodology, screening volume and uptake. To assess for risk-of-bias, we used the Joanna Briggs Institute Critical Appraisal tool.
Results Twenty-six cross-sectional descriptive studies were included out of 935 independent records. Reductions in screening volume and uptake rates were observed among eight countries. Changes in screening participation volume in five countries with national population-based screening ranged from -13% to –31%. Among two countries with limited population-based programs the decline ranged from -61% to -41%. Within the USA, population participation volumes varied ranging from +18% to -39% with suggestion of differences by insurance status (HMO, Medicare, and low-income programs). Almost all studies had high risk-of-bias due to insufficient statistical analysis and confounding factors.
Discussion and Conclusion Extent of COVID-19-induced reduction in breast screening participation volume differed by region and data suggested potential differences by healthcare setting (e.g., national health insurance vs private health care). Recovery efforts should monitor access to screening and early diagnosis to determine if prevention services need strengthening to increase coverage of marginalized groups and reduce disparities.
Introduction
Breast cancer is the most common cancer worldwide with 2.3 million women diagnosed and 685000 deaths in 2020 (WHO, 2022). Mammography-based screening programs allow for early detection of breast cancers, allowing for earlier intervention and disease stage that improves patient outcomes (IARC, 2022). Early detection and diagnosis from screening may reduce mortality up to 65% among breast cancer patients (Berry et al. 2005). Populations with a good uptake rate in screening programs can achieve a 90% 5-year survival rate in patients who received an early diagnosis attributed to screening (WHO, 2020).
The ongoing COVID-19 affected global health systems and has strained population breast mammography screening programs. Previous work on modelled evaluations and a focus on tumor staging and mortality as outcomes, suggested scenarios are likely to differ by region and organization of delivery of breast cancer screening (Figueroa et al. Prev Med 2021). In different countries, screening models vary from population-based to opportunistic screening (offered to patients in healthcare settings – more common in private healthcare) (IARC, 2016).
Here we aimed to quantify systematically breast screening participation rates before and after the COVID-19 including suspensions in nations with/without opportunistic screening programs. This was performed by investigating two primary study outcomes: changes in screening volume and participation rates.
Methods
We performed a rapid review (Tricco et al. 2015), where systematic review processes were modified to facilitate project completion within a shortened timeframe. Searches were limited to two databases and English-language governmental grey literature.
Literature search
RL ran a systematic search in on “Ovid MEDLINE(R) and In-Process, In-Data-Review & Other Non-Indexed Citations” Database and WHO COVID-19 Literature Database with entry date limits from 1 January 2020 to 12 March 2022. In brief, we performed the search with MeSH subject headers and free text terms for “COVID-19”, “Breast Neoplasms” and “Mass screening”. Our search strategies are listed in Table 1. We searched grey literature from government health websites known to have data from population-based screening programs. These consisted of the National Cancer Institute (USA), CDC (USA), NHS (National Healthcare Service) UK database, BreastScreen Australia and BreastScreen Aotearoa New Zealand. We further screened reference lists of the retrieved eligible publications to identify additional relevant studies. An English language restriction was placed on the searches. Deduplication was carried out as part of upload to Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org.
Inclusion and Exclusion Criteria
The Population, Interventions, Comparator, Outcomes, and Study Characteristics (PICOS) model (Schardt et al. 2007) was used to determine eligibility criteria. A pilot literature screen (n = 10) was performed by (RL with guidance from MD and JF) to confirm validity of criteria.
The population of focus are women eligible for breast cancer screening programs (population-based or opportunistic) globally or breast screening programs that are a part of the International Screening Cancer Network (ISCN). The intervention investigated involves the introduction of COVID-19 infection control measures.
These were assumed to be present globally due to worldwide prevalence of COVID-19 by March 2020 (chosen due to WHO’s declaration of a pandemic. The comparator involved comparing breast cancer screening statistics after COVID-19 related screening shutdown versus an analogous period in the previous years (e.g., comparing statistics in Australia from May-Sep 2020 against data from May-Sep 2018/2019) or any relevant period.
Outcomes assessed were volume of breast screening participation (“volume”) defined as total number of breast screening procedures; participation uptake rate (“uptake rate”) of breast screening program defined as the percentage of the eligible population who attend screening; and incidence of breast cancer diagnosis. These were compared between each comparator period.
Full-text, English-language primary papers or governmental published grey literature were included.
Studies with data entirely pertaining to diagnostic imaging were excluded or with future modelled data were excluded. All studies focused on women
Studies were required to have data on breast screening following the resumption of breast screening in countries with a screening shutdown.
Title, Abstract, Full text Screen
Two reviewers (RL, JF) parallelly independently reviewed titles, abstracts, and subsequently full texts based on pre-defined inclusion and exclusion criteria. Deduplication of articles and screening was performed on Covidence. Conflict resolution was performed by discussion.
Data Extraction
Data extraction for each article was conducted by a single reviewer (RL). A second reviewer (WX) then checked for eligibility of extracted data in 70% of the texts. Any conflicts were resolved by a third reviewer (JF). Data relevant to the evidence for population-based or opportunistic breast cancer screening programs during COVID-19 were extracted including citation details, publication type, study design, country, region, population, study setting, screening sample size, screening timeframe, screening volumes change (before/after COVID-19 infection control guidelines), screening participation rates change (before/after COVID-19 infection control guidelines), breast cancer incidence rates. A standardized data extraction form was created and piloted for extraction of primary outcome measures.
Risk of Bias Assessment
All studies included had cross-sectional designs. We used the Joanna Briggs Institute Critical appraisal tool for cross-sectional studies to assess the risk of bias of each article (Joanna Briggs Institute, 2022). The JBI checklist is available in Table 4. Risk-of-bias for each article was assessed by a single reviewer [RL], and a second reviewer [WX] cross-assessed the results and verified all related judgement and rationales. Discrepancies were resolved through discussion and a joint reassessment of studies.
Data Synthesis
Data were synthesized descriptively since a meta-analysis was not appropriate due to heterogeneity of data. Data was collected by comparing outcome measures before and after COVID-19 infection control measures were introduced; this was presumed due to the worldwide prevalence of COVID-19 by March 2020. This included changes in breast screening volume, participation rates and breast cancer diagnosis rates (Figueroa et al, 2021; WHO, 2020).
Data were obtained from any point after lifting of COVID-19 breast screening suspension measures until an endpoint of 31 December 2020. If quantitative data was limited, the last data point of the study was analyzed. This was compared to data from an analogous pre-COVID-19 period in 2018-2019, or if data was unavailable, against any relevant pre-pandemic period. For countries with no breast screening suspension in 2020, data from during the COVID-19 pandemic was compared with an analogous period of 2018-2019. This phenomenon only occurred in Taiwan, China (Shen et al. 2022). A percentage change against the overall comparator period was calculated.
Results
Figure 1 summarizes the search strategy. The initial search retrieved 1207 articles and 935 independent records. After screening (see Methods), 26 cross-sectional studies from 12 countries were eligible for inclusion (Table 2). Seven reports came from Europe (Campbell et al, 2021; Jidkova et al, 2022; Knoll et al, 2022; Eijkelboom et al, 2021; Losurdo et al, 2022; Toss et al, 2021; NHS England, 2021), two from Oceania (BreastScreen Australia, 2021; BreastScreen Aoteroa, 2021), one from Asia (Shen et al, 2022), two from South America (Bessa et al, 2021; Riberio et al, 2022) and 14 from North America (Chiarelli et al, 2021; Walker et al, 2021; Doubouva et al, 2021; Chen et al, 2021; Amornsiripanitch et al, 2021; Becker et al, 2021; DeGroff et al, 2021; Dennis et al, 2021; Fedewa et al, 2021; Lehman et al, 2021; London et al, 2022; Miller et al, 2021; Sprague et al, 2021; Nyante et al, 2021). The most frequently reported country was the USA (n = 11). Studies examined either regional (n = 13) or national populations (n = 13). Analysis of data from all studies was limited from 1 January 2020 to 31 December 2020.
Screening volume changes over study period
Summary data from 17 studies in eight countries reporting breast cancer screening volumes, data from 106,484,908 women before and after COVID-19 infection control measures were extracted (data from 2017 to 2020 were the comparison time period, Table 3; Doubouva et al, 2021; Bessa et al, 2021; Riberio et al, 2022; Chiarelli et al, 2021; Losurdo et al, 2022; Walker et al, 2021; NHS England, 2021; Shen et al, 2022; BreastScreen Australia, 2021; DeGroff et al, 2021; Lehman et al, 2021; Amornsiripantich et al, 2021; Sprague et al, 2021; London et al, 2022; Miller et al, 2021; Nyante et al, 2021; Becker et al, 2021). Most studies that showed calendar period trends of screening volume, noted temporal variation with declines especially at the height of the pandemic between March-May 2020. In countries with national screening programs, a negative change in screening volume was reported with the lowest volume change estimated at - 12.86% in Australia (BreastScreen Australia, 2021) followed by -31.30% in Ontario, Canada (Chiarelli et al, 2021). A larger negative change in screening volume was observed in Brazil (−41.49%) (Ribero et al, 2022) and Mexico (−61.30%) (Doubouva et al, 2021. It should be noted that Brazil and Mexico have a lower proportion of population-based breast screening coverage relative to other countries; Brazil having coverage of ∼24%, and Mexico having ∼20% coverage of the eligible population (OECD, 2021; Unger-Saldaña et al, 2020). A significant proportion of breast screening in Brazil and Mexico consists of opportunistic screening programs.
In the USA, which has mix of insurance providers there was a wide range of change in screening volume. Using data from Health Managed Organization (HMO) Blue Cross Blue Shield (BCBS) from the state of Michigan, the authors observed temporal changes in rates with an increase slightly above 2019 levels in the last few months of the year, with an 18.10% overall increase in screening volume (Becker et al. 2021). However, it’s important to note that although rates were above 2019 levels, the authors noted that the odds that a woman received breast cancer screening remained 20% lower in 2020 relative to 2019 (Becker et al. 2021). This was consistent with the decrease in screening volume that was generally observed from six studies with data among populations wholly or partially covered by national insurance (Lehman et al, 2021; Amornsiripanitch et al, 2021; Sprague et al, 2021; London et al, 2022; Miller et al, 2021; Nyante et al, 2021). Percentage decreases ranged from -36.50% (Lehman et al. 2021) to -9.80% (Miller et al. 2021). Data from the USA National Breast and Cervical Cancer Early Detection Program (NBCCEDP), which provides cancer screening services to women with low income and inadequate health insurance, reported a greater decrease (−39.00%) in volume (DeGroff et al. 2021). Two other studies had smaller populations with less certainty and wider confidence intervals with one reporting an 8% increase (Nyante et al. 2021) and the other a -10% decline (London et al. 2021). In the USA, where there is a mix of national (Medicare) and private insurance depending on age, screening volume changes were similar to other national screening programs at -36.50% (Lehman et al, 2021). In contrast, a positive increase in volume was observed among private insurance providers +30% (London et al, 2022)
Screening participation rate changes
A total of nine cross-sectional studies reported breast cancer screening participation rates and represented > 46,257,402 participants from varying calendar periods across five countries (Amornsiripantich et al, 2021; Dennis et al, 2021; Fedewa et al, 2021; Chen et al, 2021; NHS England, 2021; Campbell et al, 2021; Bessa et al, 2021; BreastScreen Aoteroa, 2021; Jidkova et al, 2022). There was considerable variability in change (Table 4), ranging from +2-8% in Scotland to -43.54% in Brazil (Campbell et al, 2021; Bessa et al, 2021). In the USA, with a mix of national and private insurance usage, there was a consistent negative change in screening participation rates (Amornsiripantich et al, 2021; Dennis et al, 2021; Fedewa et al, 2021; Chen et al, 2021).
Study quality
The quality of included studies was assessed using the JBI tool (Table 5). A weakness across most studies was failure to identify and consider confounding factors. From Table 5, twenty-five studies had no issues defining the inclusion sample. Nineteen studies were clear in defining the study setting and subjects. Studies had no issues quantifying exposure of COVID-19, although this was based on temporality since all healthcare systems globally were affected (Worldometer, 2022). All studies apart from Becker et al (2021) had no issue measuring the condition through either screening appointment attendance or insurance claims data. Most studies (65%, N=17) did not define confounding factors regarding measurement of primary outcomes. Regarding comparison of volumes of screening prior to COVID-19 and observed periods, these studies did not provide source of reduction in screening capacity (e.g. due to social distancing or participation uptake). Twenty-three studies failed to provide strategies to address confounding factors (e.g., elucidating reduction in capacity and presenting it as a proportion to overall volume).
Four studies (Bessa et al. 2021; Becker et al. 2021; London et al. 2022; Doubova et al. 2021) had unclear reasons for selection of study subjects and control groups (London et al, 2022), confounding factors that were not indicated, nor strategies included to solve this. Among these four papers, vague definition of control groups resulted in a poor comparator, resulting in unreliable outcome measures.
Twenty-three studies provided basic statistical analyses (e.g. mean, adjusted rates per population) with basic data presentation. Statistical analyses were not performed in three government papers (BreastScreen Australia. 2021; NHS England. 2021; BreastScreen Aotearoa. 2021). Twenty-two studies were unclear or did not provide sufficient descriptive statistical analyses regarding comparison of control data to observed data. Statistical analyses were performed in four studies. This includes provision of odds ratios by Doubova et al (2021) and Miller et al (2021), Poisson estimation of a 95% confidence interval by Sprague et al (2021) and 95% confidence intervals from comparison of means from Nyante et al (2021).
Discussion
In this rapid review we show that during the COVID-19 pandemic, there was a generally reported reduction in breast cancer screening volume and participation that varied by healthcare setting (e.g., national population-based screening vs opportunistic or private health care). Our data suggests that participation is a major contributor, which requires monitoring by health systems and could inform prevention and early diagnosis efforts, especially if certain groups are not participating.
Fear of COVID-19 infection is likely an important source that would reduce screening participation and volumes. In populations with national population-based screening, there were competing risks between contracting COVID-19 and breast screening health benefits (Chiarelli et al. 2021; Walker et al. 2021). Considering initial lack of free access to vaccines in 2020, severity of the first COVID-19 wave, and older age of breast screening patients, patients likely translated to decreased screening participation due to infection fears (Losurdo et al. 2022). Governmental databases in England and New Zealand (NHS England, 2021; BreastScreen Aotearoa, 2021) corroborate this; participation uptake rates decreased by - 6.70% (New Zealand) to -11.80% (England). Given that uptake rate measures the proportion of patients who fulfil their appointment, monitoring this outcome could provide early indications on reductions in willingness of patients to attend screening. Shen et al (2022) describes changes in patient behaviour in Taiwan where they shifted mammography appointments from hospitals to mobile screening centres where perceived risks of contracting COVID-19 were lower. This led to greater decreases in hospital-based mammography volumes where a -41.43% reduction was observed relative to a -23.99% reduction in mobile centres. Interestingly, this phenomenon occurred despite low COVID-19 prevalence in Taiwan, China.
Reductions in screening capacity are likely another source for screening volume reductions. Screening capacity reductions were caused by social distancing, staggered appointments, staff exposure to COVID-19, and cleaning measures. This likely resulted in reductions in time allocated for screening to occur (Walker et al. 2021; Sprague et al. 2020). Sprague et al (2021) considered screening capacity when assessing screening volume. Even though screening capacity recovered to pre-pandemic levels in July 2020, screening volume experienced a 10.8% decrease relative to the control period. Reductions in screening capacity were potentially not the sole factor to screening volume reductions. However, most publications included in our rapid review did not collect data regarding screening capacity so we cannot determine the proportion of change in screening volume that was attributed to either reduction in screening capacity or change in patient willingness to attend screening. Future analyses are needed where both of these measures are obtained, which would inform what measures are needed (e.g. information campaigns to alleviate patient fears or increase clinical staffing for catch-up of missed appointments).
Our data supports differences by healthcare system which were particularly evident in data from the USA where there is a mix of private and national healthcare (Medicare for persons 65+ (Medicare.gov). DeGroff et al (2021), who studied populations reliant solely on national health insurance, showed larger screening volume reductions (−39.00%). This was relative to studies focusing solely on populations with private insurances, or studies including patients from both groups (−36.50% to +30%). Amornsiripanitch et al (2021), which included national and private insurance patients, corroborates this. Medicaid and Medicare patients had - 17.06% screening volume reduction compared to -10.50% experienced by the entire population. Miller et al (2021) suggests opportunity-cost of attending breast screening in lower income groups (e.g., employment), may have led to decreased breast screening in such populations. Some literature showed increases in screening volumes (Nyante et al. 2021; Becker et al. 2021) and uptake rates (Campbell et al. 2021). Increased volume (+9% 95% CI [-12, 32]) from Nyante et al (2021) was inconclusive due to wide 95% confidence intervals. Although this study was robust limited data collection till September 2020 did not show full extent of change regarding screening volumes after lifting of COVID-19 suspension guidelines in June 2020. From trends explored in study, breast screening rates were possibly recovering in the study population (USA) in late-2020, but more data is required. The Affordable Care Act may have alleviated breast screening cost through health insurance coverage reforms (Zhao et al. 2020). However, this does not address other underlying socio-economic inequalities (e.g. high cost of treatment, time off from work due to sickness).
Patients from deprived backgrounds may be fearful of dealing with the consequences of abnormal screening results (e.g., treatment). This may strain patient finances worsened by COVID-19, potentially explaining lower screening volumes and uptake. Future data on patient characteristics including insurance status, socioeconomic and race/ethnicity could inform targeted campaigns to reduce inequities if disparities exist.
Becker et al (2021) showed a screening volume increase after lifting of COVID-19 suspension guidelines. This study focused on patients who utilize solely private insurance. Patients already paying for services may be more inclined to maximize utilization of coverage. However, this study states that the odds that a woman received breast cancer screening remained 20% lower in 2020 (OR = 0.80 (0.80, 0.81)) relative to 2019. This study scored poorly in the JBI appraisal tool due to poor outcome measurement; it was unclear how odds ratio was derived, therefore, increasing the risk of bias of this study. Unusual outcome measures were used, that being the claims invoice for the service. This appeared unreliable; it was unclear whether paying for the service equates to a fulfilled appointment. Invoices could be delayed, making it unclear when the screening took place. This study’s evidence quality needs to be increased for results to be conclusive.
Campbell et al (2021) states a 10.96% increase in uptake rate in Scotland. This study population (within study period) solely included patients who had their appointments cancelled in March 2020 due to the 1st lockdown, and high-risk patients. This particular patient group may have an increased urgency to catch up on screening. This could have contributed to the increased uptake rate of screening in Scotland in the study period. Increase in uptake rates could also be attributed to increased accessibility for patients due to the “work-from-home” model and increased health consciousness due to COVID-19. Neither raw data nor sample size was defined in study and will require future analysis.
Due to the inherent weaknesses of a rapid review, certain limitations are present within the study as explored below. However, this study can be expanded upon by various means (also explored below) to further elucidate the global impact of COVID-19 on breast cancer detection and subsequent care. Other limitations include the COVID-19 pandemic context as an evolving field with fast publication turnovers; more papers could have been published since the review started. This issue could be partially addressed by completing a repeat search with employment of forwards and backwards citation tracking, while including more grey literature sources apart from governmental databases (e.g., private screening databases).
Other limitations included studies had insufficient data for combined analysis regarding COVID-19 waves past December 2020. Additionally, data obtained was cross-sectional instead of cohort-based; we were unable to analyze trends and recovery in breast cancer screening rates and incidence rates over time. Exclusion of non-English language literature was a weakness. Many countries with extensive population-based breast screening programs that were affected by COVID-19 in Europe and Asia were unaccounted for; the inclusion of additional data would be useful to clarify the impact of the pandemic on breast cancer screening program uptake.
In summary, screening volume and uptake rates were generally reduced but many studies showed gains over time even if overall a decline in screening volume observed. These declines were likely due to the first COVID-19 wave where many health care facilities paused non-essential services. Volume and uptake reductions of smaller magnitudes were observed and our data suggest some difference depending on region and health care coverage. Access to screening services may increase marginalization of some vulnerable groups in the USA due to the pandemic and recovery efforts to reduce disparities in access to screening and early diagnosis should be monitored to determine if prevention services need strengthening. Participation and volume are not conclusive endpoints themselves and future work on from registries and other data sources are needed to determine if there has been any impact on incidence, stage and mortality outcomes.
Data Availability
All data analysed is provided in the manuscript or supplements
Footnotes
Funding: There was no source of funding for this project.
ET is supported by a Cancer Research UK Career Development Fellowship (C31250/A22804)